CN110766672B - Electrode fouling detection method - Google Patents
Electrode fouling detection method Download PDFInfo
- Publication number
- CN110766672B CN110766672B CN201911001832.4A CN201911001832A CN110766672B CN 110766672 B CN110766672 B CN 110766672B CN 201911001832 A CN201911001832 A CN 201911001832A CN 110766672 B CN110766672 B CN 110766672B
- Authority
- CN
- China
- Prior art keywords
- electrode
- image
- self
- graph data
- quotient graph
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/187—Segmentation; Edge detection involving region growing; involving region merging; involving connected component labelling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
- G06V10/75—Organisation of the matching processes, e.g. simultaneous or sequential comparisons of image or video features; Coarse-fine approaches, e.g. multi-scale approaches; using context analysis; Selection of dictionaries
- G06V10/751—Comparing pixel values or logical combinations thereof, or feature values having positional relevance, e.g. template matching
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Computing Systems (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Databases & Information Systems (AREA)
- Artificial Intelligence (AREA)
- Health & Medical Sciences (AREA)
- Multimedia (AREA)
- Quality & Reliability (AREA)
- Image Processing (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The invention discloses an electrode fouling detection method. The method comprises the following steps: preprocessing the electrode image to obtain an electrode gray image; performing opening operation and closing operation on the electrode gray level image according to templates in the morphological operator template set to obtain an electrode foreground image and an electrode background image; carrying out self-quotient graph operation by utilizing the electrode foreground image and the electrode background image to obtain initial self-quotient graph data; establishing an optimal self-quotient graph data selection model, acquiring optimal self-quotient graph data, and performing dispersion standardization processing on the optimal self-quotient graph data; performing thresholding treatment on the normalized self-quotient graph data to obtain a connected domain image; and extracting an outer layer contour of the connected domain image, filling to obtain mask data, and performing XOR operation on the mask data and the connected domain image to obtain a smudge segmentation result. The method can realize the dirty segmentation of the electrode, has strong adaptability to illumination change, reduces the workload of manual parameter adjustment, and has high calculation efficiency and high detection precision.
Description
Technical Field
The invention relates to the technical field of electrode detection, in particular to a defect detection method of an electrode.
Background
An electrode is a sensor element that is comb-shaped across the electrode, typically with a square wave passing across the electrodes for reactance-based calculations. The electrodes need to be inspected and cleaned during production, but there are cases where the electrodes are not well encapsulated in use, such as glass substrates and the like. Since general dirt can absorb part of electromagnetic waves, if the electrode is in contact with the dirt without any protective measures, problems such as short circuit and signal attenuation occur.
Currently, methods for detecting fouling include two. One is designed around the MURA (display unevenness) inspection requirements of products such as liquid crystal panels, and it is necessary to provide environments with special requirements for light sources, such as darkrooms, collimated light sources, and stripe light sources. Because the electrode size is tiny, and the locating position is various, and some circumstances need on-line monitoring or detect with the help of the instrument when maintaining. The electrode defect detection is not only a production process but also a common equipment self-maintenance process, so that the detection equipment needs to be an independent image processing module or even a handheld equipment, and the light source is directly uncontrollable, thereby influencing the precision of the MURA detection method.
Another type of smear detection is based on discrete fourier transform DFT and designed traps to extract features. The detection method comprises the steps of firstly calculating a frequency spectrum, and then manually designing a notch mask by visually observing response bright spots of the frequency spectrum. The process has excessive fixed parameters and needs frequent debugging, and for electrode detection, the significance of taking and designing a special algorithm specially according to various working conditions is not large. General detection equipment is directly taken pictures, so that collected images are mostly different, manual parameters such as a notch position and the like need to be adjusted again, and if the input size of the images changes slightly, the notch position is likely to deviate from the response position of a frequency spectrum, so that other regular textures are introduced. In addition, DFT has a certain limitation on the image of the thick line strong edge such as an electrode, and a scene using a DFT equal frequency domain filtering method generally has a texture with a lower amplitude and a higher frequency. For comb electrodes and the like, due to the existence of obvious step-type boundaries, the brightness change is large, frequency positions corresponding to textures are difficult to find in the DFT spectrum, and all relevant frequency responses are difficult to completely eliminate. Moreover, the gradient of the electrode changes more according to different model directions, so that the trap positions are different. Therefore, it is difficult to process using a method of manually designing a trap.
The MURA and DFT methods based on dedicated light sources limit the feasibility of detection, subject to the manner of use. Therefore, the existing electrode element dirt detection technology has the problems of high requirement on a light source, large workload of manual parameter adjustment of a feature extraction method, large calculation amount of the detection method and low detection precision.
Disclosure of Invention
The invention provides an electrode dirt detection method, which realizes electrode dirt segmentation, has strong adaptability to illumination change, reduces the manual parameter adjustment workload, and has high calculation efficiency and high detection precision.
A method of electrode fouling detection, the method comprising:
firstly, preprocessing an electrode image obtained by a camera to obtain an electrode gray image;
step two, constructing a morphological operator template set, and sequentially performing opening operation and closing operation on the electrode gray level image according to the templates in the set to obtain an electrode foreground image and an electrode background image corresponding to the morphological operator template;
performing self-quotient graph operation by using the electrode foreground image and the electrode background image corresponding to the morphological operator template, and eliminating uneven illumination to obtain initial self-quotient graph data;
establishing an optimal self-quotient graph data selection model, processing the initial self-quotient graph data to obtain optimal self-quotient graph data, and performing dispersion standardization processing on the optimal self-quotient graph data to obtain electrode self-quotient graph data;
fifthly, carrying out thresholding treatment on the electrode self-quotient graph data, carrying out connected domain detection analysis on a thresholding result, and obtaining a connected domain image;
and sixthly, carrying out contour analysis on the connected domain image, extracting an outer contour and filling to obtain mask data, and carrying out XOR operation on the mask data and the connected domain image to obtain a dirt segmentation result.
The first step is specifically as follows: the electrode image acquired by the camera is a color image, the preprocessing is graying processing, and the graying method specifically comprises the following steps: according to RGB data of the electrode image, obtaining an electrode gray image I with strong contrast by using a dark channel prior mode: i (x, y) ═ min (R (x, y), G (x, y), B (x, y)), where I (x, y) is the value of the electrode gray-scale image I at position (x, y), and R (x, y), G (x, y), B (x, y) represent the red, green, and blue channel component values of the pixel at position (x, y), respectively.
The morphology operator template set comprises a cross template, an isotropic template and a cavity template.
The self-quotient graph operation specifically comprises the following steps:
wherein, F (x, y), B (x, y) are the values of the electrode foreground image and the electrode background image at the position (x, y), respectively, and SQI (x, y) is the value of the initial self-quotient graph data at the position (x, y).
The best self-quotient graph data selection model is as follows: performing median calculation on the initial self-quotient graph data; and selecting the initial self-quotient graph data with the maximum median value as the optimal self-quotient graph data.
The threshold value determination method of the thresholding comprises the following steps: and determining a segmentation threshold value by adopting a maximum inter-class variance method.
The invention has the beneficial effects that:
1. according to the method, the obtained electrode foreground image and the electrode background image are subjected to self-quotient graph calculation, the influence of uneven illumination is eliminated, and the adaptability to illumination change is strong;
2. according to the method, the electrode foreground image and the electrode background image are obtained by utilizing the morphological operator template set, the characteristic extraction parameters do not need to be manually designed, and the manual parameter adjustment workload is reduced;
3. in the invention, the calculation of the self-quotient graph is the only floating point type calculation step, so that the calculation speed is improved, the calculation time is reduced, and the calculation efficiency is high;
4. the method can acquire the optimal self-quotient graph data by utilizing the morphological operator template set and the optimal self-quotient graph data selection model, and has high detection precision.
Drawings
FIG. 1 is a diagram of an electrode fouling detection process;
FIG. 2 is a schematic diagram of an electrode foreground image and an electrode background image obtained under each morphological operator template;
fig. 3 is a graph showing the results of electrode contamination division.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
The invention provides an electrode fouling detection method. The invention is suitable for an area-array camera, extracts brightness information of a captured electrode image, performs filtering through a morphological operator template set, further extracts an electrode region and a peripheral region, and improves response precision through a self-quotient image mode, thereby providing an accurate segmentation result. The following description will be made by way of specific examples. The electrode fouling detection process diagram is shown in fig. 1.
The first embodiment is as follows:
the electrode fouling detection method comprises the following steps:
step one, preprocessing an electrode image obtained by a camera to obtain an electrode gray image.
The image is preprocessed, and if the image is a color image, the colors of the electrode part are FPC yellow and metal silver according to experience, and the background is black or green, red, yellow and blue of the circuit board. Therefore, when a color image is encountered, the minimum value of the corresponding position component of three channels is taken by using a dark channel prior mode, and an image I with stronger contrast is obtained:
I(x,y)＝min(R(x,y),G(x,y),B(x,y))
wherein I (x, y) is the value of the electrode gray image I at the position (x, y), and R (x, y), G (x, y), B (x, y) represent the red, green, blue channel component values of the pixel at the position (x, y), respectively
And step two, constructing a morphological operator template set, and sequentially performing opening operation and closing operation on the electrode gray level image according to the templates in the set to obtain an electrode foreground image and an electrode background image corresponding to the morphological operator template.
The morphological operator template set comprises cross templates K with different sizes in consideration of symmetry in electrode designXIsotropic template KRHollow form KQ. Taking template size 3x3 as an example, for a grid-type electrode, there is a cross template:
for V-shaped electrodes, there is an isotropic template:
for serpentine electrodes, there is a cavity template:
the template size may be 3x3, 5x 5. Further, templates of template size 7x7 may also be added, and so on.
The method for sequentially carrying out opening operation and closing operation on the electrode gray level image comprises the following steps: the inner circulation carries out opening and closing operation on the electrode gray level image in sequence according to the sequence of the cross template, the uniform template in each direction and the cavity template, and the outer circulation carries out opening and closing operation on the electrode gray level image in sequence according to the sequence of the sizes of the templates from small to large so as to obtain an electrode foreground image and an electrode background image.
The method is implemented by completing the opening operation and the closing operation of an image based on a maximum value filter and a minimum value filter, and comprises the following steps:
with 3x3 template
A maximum filter is used as an example. For a certain pixel position (x, y), KRAround which a window of size 3X3 is formed. At KRSampling pixels in the neighborhood of 3x3 with (x, y) as the center in the part which is not zero, and sorting the obtained pixel values to obtain the maximum value max (I (x, y) × K)R) And a minimum value min (I (x, y) × KR) Wherein the maximum value is the maximum value filtered pixel value I of the pixel positionMax(x, y), the minimum value is the minimum value filtered pixel value I of the pixel positionMin(x,y)：
IMax(x,y)＝max(I(x,y)*KR)＝max(I(x-1,y-1),……,I(x+1,y+1))
IMin(x,y)＝min(I(x,y)*KR)＝min(I(x-1,y-1),……,I(x+1,y+1))
The on operation and the off operation are realized by controlling the order of the maximum value filtering and the minimum value filtering, and the original position of the contour can be maintained. The on operation is to perform minimum filtering first and then maximum filtering, and the effect is to close the local maximum part. And the closed operation firstly carries out maximum value filtering and then carries out minimum value filtering, and the opposite effect is realized. An electrode background image B is obtained through an opening operation, and an electrode foreground image F, namely an electrode area, is obtained through a closing operation. The result of the opening and closing operation is shown in fig. 1.
And thirdly, carrying out self-quotient graph operation by using the electrode foreground image and the electrode background image corresponding to the morphological operator template, and eliminating uneven illumination to obtain initial self-quotient graph data.
The feature of the morphological filtering for the electrode is a relatively suitable method, but the method using the morphological filtering also has the problem of manual design because factors such as a visual angle, an electrode interval and the like need to be continuously adjusted to determine parameters. Meanwhile, the morphological filtering method is difficult to eliminate the phenomenon of uneven illumination when the image is collected.
The SQI is an abbreviation of a self-quotient image, and the numerator image and the denominator image are divided point by point, so that the SQI is a flexible high-pass filter. The data for SQI has the following characteristics: when both the numerator and the denominator carry low-frequency information, the amplitude of the low-frequency information is infinitely close to 1 due to division and difference, and the high frequency is generally the difference between the numerator and the denominator and is far away from 1 in a multiple relation. Therefore, the SQI can eliminate the illumination unevenness while enhancing the high frequency characteristics. The SQI describes high-frequency information by multiples, can prevent the problem of excessive gray scale loss during integer quantization, and does not need to design a gray scale mapping curve manually, such as gamma conversion and the like. The obtained SQI image can be directly subjected to thresholding processing.
The initial self-quotient graph data obtained from the electrode foreground image and the electrode background image corresponding to each morphological operator template is shown in fig. 2, the template size is represented in a transverse direction, and the sizes are 3X3 and 5X5 respectively; the longitudinal direction represents a cross template, an isotropic template, and a void template. The calculation process of the initial quotient graph data SQI is as follows: using SIMD instructions as an example, an implementer may perform RSQRT calculations with the electrode foreground image as a denominator and the electrode background image as a numerator. First, point-by-point division is performed, and then RSQRT is performed on the result. Acceleration by SIMD is possible.
F (x, y), B (x, y) are the values of the electrode foreground image and the electrode background image at the position (x, y), respectively, and SQI (x, y) is the value of the initial self-quotient graph data at the position (x, y). RSQRT is reciprocal and square, i.e.
The SIMD instruction set content which is prepared by RISC and CISC can realize the amplitude adjustment effect of the evolution in the same instruction cycle. The significance of the evolution here is to improve the contrast of the soiling. The practitioner can also use the SQI equation as is:
since the SQI already stretches contrast, the implementer can choose not to proceed according to the real situation or perform the processing according to the camera characteristics or the dirty characteristics by using other incremental mapping functions, such as
And the like. The implementer can choose the implementation method by the above equation, but the core content of the surrounding is the calculation of a self-quotient graph by dividing the electrode background image and the electrode foreground image. The step is the only floating point calculation step of the invention, and the calculation efficiency of the method is improved on the whole.
And step four, establishing an optimal self-quotient graph data selection model, processing the initial self-quotient graph data to obtain optimal self-quotient graph data, and performing dispersion standardization processing on the optimal self-quotient graph data to obtain electrode self-quotient graph data.
Since the problem of uneven illumination is eliminated, the gray level of the bright portion can be represented by a median value. Performing median calculation on the initial self-quotient graph data SQI, namely sorting the pixel gray scale of the whole image, and taking a median M:
M＝MEDIAN(sort(SQI))
where sort () is a sorting function and media () is used to find the MEDIAN of the sorting result. When the number of image pixels is odd, media () returns the value of the middle element; when the number of image pixels is even, media () returns the average of the middle two elements. And D, respectively calculating the median M corresponding to each template according to the initial self-quotient graph data of each template obtained in the step three.
Referring to the line-by-line variation of fig. 2, the extracted foreground and background purity is different due to the different regular texture responses, and therefore the intensity of the SQI is different. The morphological filtering is nonlinear filtering based on logic judgment, and in the process of selecting the optimal structure, once the morphological operator template can effectively extract features, the response of the SQI becomes a flat bright area, namely the difference between the current median and the median of the previous option can be several times. When the texture is matched with the template, the median value is obviously increased, the difference of corresponding regions is extremely large, and an SQI image with high quality can be obtained. Moreover, once the morphology template matches the texture of the electrode voids, the median value of the image hardly changes. Starting from the sparsest, smallest template, the selected structure is considered to be optimal when the resulting median is the largest.
The best self-quotient graph data selection model is as follows: when the next M' is smaller than the current M, the current structure is considered to be the best structure. When the next M' is larger than the current M and below the threshold T-5, the search is stopped and the current K is considered to be the best option. The threshold T is an empirical value, and can be used as an option for controlling the image quality, and if the threshold is higher, the stability is better, but the calculation amount is larger. Therefore, the self-quotient graph data corresponding to the maximum median value, namely the best self-quotient graph data, can be obtained.
And the optimal self-quotient graph data is subjected to dispersion standardization, so that the image can be conveniently quantized to an integer range. The distribution of the SQI image is around 1, so to match the [0,1] range of the visualization, the entire image is dispersion normalized: SQI ═ minmaxnormalize (SQI). Thus, electrode self-quotient graph data is obtained.
And fifthly, carrying out thresholding treatment on the electrode self-quotient graph data, carrying out connected domain detection analysis on a thresholding result, and obtaining a connected domain image.
And after the electrode self-quotient graph data SQI is quantized to an integer range, carrying out maximum inter-class variance OTSU thresholding on the electrode self-quotient graph data SQI. At this time SQI, the dirty feature is significantly magnified and the bright flat area approaches the median. The histogram is sharp in peak and robust to automatic thresholding.
And sixthly, carrying out contour analysis on the connected domain image, extracting an outer contour and filling to obtain mask data, and carrying out XOR operation on the mask data and the connected domain image to obtain a dirt segmentation result.
And carrying out contour analysis on the connected domain image CONNEC, and extracting the outermost layer contour of the connected domain. The MASK is obtained by filling the interior of the contour with a solid color, the effect of filling FILL UP being shown in fig. 1. And performing exclusive or operation on the MASK data MASK and the connected domain image CONNEC point by point:
thus, a dirty segmentation result is obtained. Fig. 3 shows an electrode image, self-quotient graph data, and a dirty segmentation result graph from left to right.
The processing method extracted by the invention has lower calculation amount, and is convenient for an implementer to use the SIMD instruction set possessed by most processors for acceleration. The implementer can use the method of the invention to integrate the detection function into a handheld device or a low-power consumption instrument without using a special light source.
The above embodiments are merely preferred embodiments of the present invention, which should not be construed as limiting the present invention, and any modifications, equivalents and improvements made within the spirit and principle of the present invention should be included in the scope of the present invention.
Claims (3)
1. A method of detecting fouling of an electrode, the method comprising:
firstly, preprocessing an electrode image obtained by a camera to obtain an electrode gray image;
step two, constructing a morphological operator template set, wherein the morphological operator template set comprises cross templates, all-directional consistent templates and cavity templates with different sizes, according to the templates in the set, performing open operation on the electrode gray level image to obtain an electrode background image corresponding to the morphological operator template by an inner loop in the sequence of the cross templates, all-directional consistent templates and the cavity templates and performing closed operation on the electrode gray level image to obtain an electrode foreground image corresponding to the morphological operator template by an outer loop in the sequence of the sizes of the templates from small to large;
performing self-quotient graph operation by using the electrode foreground image and the electrode background image corresponding to the morphological operator template, and eliminating uneven illumination to obtain initial self-quotient graph data; the self-quotient graph operation specifically comprises the following steps:
wherein, F (x, y), B (x, y) are values of the electrode foreground image and the electrode background image at the position (x, y), respectively, and SQI (x, y) is a value of the initial self-quotient graph data at the position (x, y);
step four, establishing an optimal self-quotient graph data selection model, wherein the optimal self-quotient graph data selection model is as follows: performing median calculation on the initial self-quotient graph data; selecting initial self-quotient graph data with the maximum median value as optimal self-quotient graph data; carrying out dispersion standardization processing on the optimal self-quotient graph data to obtain electrode self-quotient graph data;
fifthly, carrying out thresholding treatment on the electrode self-quotient graph data, carrying out connected domain detection analysis on a thresholding result, and obtaining a connected domain image;
and sixthly, carrying out contour analysis on the connected domain image, extracting an outer contour and filling to obtain mask data, and carrying out XOR operation on the mask data and the connected domain image to obtain a dirt segmentation result.
2. The method according to claim 1, wherein the first step is specifically: the electrode image acquired by the camera is a color image, the preprocessing is graying processing, and the graying method specifically comprises the following steps: according to RGB data of the electrode image, obtaining an electrode gray image I with strong contrast by using a dark channel prior mode: i (x, y) ═ min (R (x, y), G (x, y), B (x, y)), where I (x, y) is the value of the electrode gray-scale image I at position (x, y), and R (x, y), G (x, y), B (x, y) represent the red, green, and blue channel component values of the pixel at position (x, y), respectively.
3. The electrode contamination detection method according to claim 1, wherein the thresholding process is performed by a threshold determination method comprising: and determining a segmentation threshold value by adopting a maximum inter-class variance method.
Priority Applications (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010702878.5A CN111833339A (en) | 2019-10-21 | 2019-10-21 | Electrode fouling detection method |
| CN201911001832.4A CN110766672B (en) | 2019-10-21 | 2019-10-21 | Electrode fouling detection method |
| CN202010702877.0A CN111833338A (en) | 2019-10-21 | 2019-10-21 | Electrode fouling detection method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911001832.4A CN110766672B (en) | 2019-10-21 | 2019-10-21 | Electrode fouling detection method |
Related Child Applications (2)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010702877.0A Division CN111833338A (en) | 2019-10-21 | 2019-10-21 | Electrode fouling detection method |
| CN202010702878.5A Division CN111833339A (en) | 2019-10-21 | 2019-10-21 | Electrode fouling detection method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110766672A CN110766672A (en) | 2020-02-07 |
| CN110766672B true CN110766672B (en) | 2020-11-20 |
Family
ID=69331520
Family Applications (3)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010702877.0A Withdrawn CN111833338A (en) | 2019-10-21 | 2019-10-21 | Electrode fouling detection method |
| CN202010702878.5A Withdrawn CN111833339A (en) | 2019-10-21 | 2019-10-21 | Electrode fouling detection method |
| CN201911001832.4A Active CN110766672B (en) | 2019-10-21 | 2019-10-21 | Electrode fouling detection method |
Family Applications Before (2)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010702877.0A Withdrawn CN111833338A (en) | 2019-10-21 | 2019-10-21 | Electrode fouling detection method |
| CN202010702878.5A Withdrawn CN111833339A (en) | 2019-10-21 | 2019-10-21 | Electrode fouling detection method |
Country Status (1)
| Country | Link |
|---|---|
| CN (3) | CN111833338A (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7159758B1 (en) * | 2003-06-26 | 2007-01-09 | Emc Corporation | Circuit board processing techniques using solder fusing |
| CN105181706A (en) * | 2015-09-23 | 2015-12-23 | 电子科技大学 | Bad defect detection method for SMD resistor on substrate |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7075642B2 (en) * | 2003-02-24 | 2006-07-11 | Intel Corporation | Method, structure, and apparatus for Raman spectroscopy |
| CN103872983B (en) * | 2014-04-04 | 2016-05-25 | 天津市鑫鼎源科技发展有限公司 | Solar cell surface defect inspection method |
| CN104102903A (en) * | 2014-07-14 | 2014-10-15 | 杭州电子科技大学 | Secondary face recognition method based on SRC |
| CN108830838A (en) * | 2018-05-28 | 2018-11-16 | 江苏大学 | A kind of pcb board incompleteness Trigger jitter detection method of sub-pixel |
| CN110012287B (en) * | 2019-02-16 | 2019-11-05 | 深圳市创生达电子有限公司 | The dirty self checking method of digital camera image sensor based on retina perception |
| CN110132978A (en) * | 2019-04-28 | 2019-08-16 | 无锡和博永新科技有限公司 | Resistor disc full wafer detection device and detection method |
-
2019
- 2019-10-21 CN CN202010702877.0A patent/CN111833338A/en not_active Withdrawn
- 2019-10-21 CN CN202010702878.5A patent/CN111833339A/en not_active Withdrawn
- 2019-10-21 CN CN201911001832.4A patent/CN110766672B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7159758B1 (en) * | 2003-06-26 | 2007-01-09 | Emc Corporation | Circuit board processing techniques using solder fusing |
| CN105181706A (en) * | 2015-09-23 | 2015-12-23 | 电子科技大学 | Bad defect detection method for SMD resistor on substrate |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111833338A (en) | 2020-10-27 |
| CN110766672A (en) | 2020-02-07 |
| CN111833339A (en) | 2020-10-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108921176B (en) | Pointer instrument positioning and identifying method based on machine vision | |
| CN107507173B (en) | No-reference definition evaluation method and system for full-slice image | |
| CN108460757B (en) | Mobile phone TFT-LCD screen Mura defect online automatic detection method | |
| CN107316077B (en) | Automatic adipose cell counting method based on image segmentation and edge detection | |
| CN112614062B (en) | Colony counting method, colony counting device and computer storage medium | |
| CN108876768B (en) | Shadow defect detection method for light guide plate | |
| CN110378902B (en) | Scratch detection method under high noise background | |
| CN107123124B (en) | Retina image analysis method and device and computing equipment | |
| Bibiloni et al. | A real-time fuzzy morphological algorithm for retinal vessel segmentation | |
| CN115205223B (en) | Visual inspection method and device for transparent object, computer equipment and medium | |
| CN108765406A (en) | A kind of snow mountain detection method based on infrared remote sensing image | |
| CN113592782B (en) | Method and system for extracting X-ray image defects of composite material carbon fiber core rod | |
| CN109064439B (en) | Partition-based single-side light-entering type light guide plate shadow defect extraction method | |
| CN116664457B (en) | Image processing method for enhancing denoising | |
| CN114998664A (en) | Rapid detection method and device for micro-plastic in seawater by multiple optical platforms | |
| CN114037691A (en) | Carbon fiber plate crack detection method based on image processing | |
| CN108830851B (en) | LCD rough spot defect detection method | |
| Guarneri et al. | 3D remote colorimetry and watershed segmentation techniques for fresco and artwork decay monitoring and preservation | |
| Ma et al. | An automatic detection method of Mura defects for liquid crystal display | |
| CN113008524B (en) | Screen light leakage detection method and device | |
| Pollatou | An automated method for removal of striping artifacts in fluorescent whole-slide microscopy | |
| CN111738984B (en) | Skin image spot evaluation method and system based on watershed and seed filling | |
| CN110766672B (en) | Electrode fouling detection method | |
| JP3854585B2 (en) | Display defect detection method and display defect inspection apparatus for liquid crystal panel | |
| CN115131336B (en) | Dark line defect detection method under four-color background picture of display screen |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20201028 Address after: Room 403, North R & D building, Central Europe Industrial Park, Jingba Road, Dongtai Economic Development Zone, Yancheng City, Jiangsu Province 224200 Applicant after: Yancheng Zhaoyang Industrial Design Co., Ltd Address before: No. 3A16 and 3a17, building 1, incubation, University Science Park, No. 11, Changchun Road, high tech Industrial Development Zone, Zhongyuan District, Zhengzhou City, Henan Province Applicant before: Zhengzhou Maitou Information Technology Co.,Ltd. |
|
| TA01 | Transfer of patent application right | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |